Bug 1729909

Description of problem:

When using nvme disks in aws, it appears the kubelet is not able to retrieve the partition type:

- The first time the volume is attached to a node it gets partitioned, it's formatted and mounted.

- Subsequents tries to attach and mount the volume fails because the partition can't be detected. It ends failing with:
=====
failed to mount the volume as \"xfs\", it already contains unknown data, probably partitions. Mount error: mount failed: exit status 32\nMounting command: systemd-run\nMounting arguments: --description=Kubernetes transient mount for /var/lib/origin/openshift.local.volumes/plugins/kubernetes.io/aws-ebs/mounts/aws/eu-west-1b/vol-00888f74479ee1602 --scope -- mount -t xfs -o defaults /dev/nvme2n1 /var/lib/origin/openshift.local.volumes/plugins/kubernetes.io/aws-ebs/mounts/aws/eu-west-1b/vol-00888f74479ee1602\nOutput: Running scope as unit run-17520.scope.\nmount: wrong fs type, bad option, bad superblock on /dev/nvme2n1
=====

Before this happens, the kubelet is not able to retrieve the partition type (dos):
=====
mount_linux.go:549] Attempting to determine if disk "/dev/nvme2n1" is formatted using blkid with args: ([-p -s TYPE -s PTTYPE -o export /dev/nvme2n1])
mount_linux.go:552] Output: "DEVNAME=/dev/nvme2n1\nPTTYPE=dos\n", err: <nil>
mount_linux.go:590] Disk dos detected partition table type: %!s(MISSING)
=====

A blkid shows the correct output:

# blkid -p -s TYPE -s PTTYPE -o export /dev/nvme2n1
DEVNAME=/dev/nvme2n1
PTTYPE=dos

The correct partition to mount should be nvme2n1p1:
nvme2n1     259:5    0  75G  0 disk 
`-nvme2n1p1 259:6    0  75G  0 part

I'm attaching full logs, pvc, pv, etc.

Version-Release number of selected component (if applicable):
3.11.98

How reproducible:
N/A - I don't have an environment to reproduce this